Predicting Stock Market Returns

Amongst the nice things about blogging is that you get to write your own headlines. I imagine that Jason Zweig isn’t all that happy with his editor’s “hed” for his column today in the Wall Street Journal. It reads “Why Market Forecasts Keep Missing the Mark.” I was a little disappointed, but not really surprised, to find that the column doesn’t really address that good question at all.

So I will try. How hard could it be?

Market forecasts, such as “the Dow will be up 14% in 2009″ are doomed to failure because the market is impossible to forecast.

That was easy.

Some things are forecastable. The weather, for example. Or how many votes a candidate will get in an election. You can look at certain data, do a little math, account for a special factor or two, and presto, a useful estimate of the near future.

But for other things you really can’t create a useful estimate. The score in the next Super Bowl. The next roll of the dice. It is not that you are stupid or lack data. And it is not that you don’t understand what is going on. You can say some useful things about these unforecastable future events. The Steelers are heavily favored. Seven is the most likely dice roll, and fourteen just ain’t gonna happen. But these are not predictions, they are statements about likelihoods and probabilities.

So it is with the stock market. (And for that matter, the bond market, commodity markets, etc.) Occasionally somebody will throw out a prediction like “the Dow will be up 14%” but nobody, the speaker included, expects it to be taken very seriously.

Sober predictions about the stock market are really estimates of the so-called expected outcome, the probability weighted average of all possible outcomes. E.g. seven is the expected outcome of a roll of two dice. The most common way to come up with an expected outcome for a year in the stock market is to average historical performance. Depending on which years are used, and which indexes, the answer is usually something like 10-12%.

So a personal finance pundit will tell you to expect 10% average annual returns in the stock portion of your investment portfolio. That’s a reasonable thing to say, but I think that too many people, possibly including the pundits, misunderstand what it means.

Imagine that on January 1, 2006, based on being told to expect 10% annual returns in the stock market, you invested $100. With the help of Microsoft Excel, you work out that you should have $672.75 on January 1, 2026. But over the next three years the investment actually goes down, so your January 1, 2009 balance is $81.88. Now what is your expected 2026 value?

Way too many people would say $672.75 or something like it. They think that the 10% number they were given was a forecast of what was actually going to happen over twenty years, rather than an estimate of the expected outcome for each year. Put another way, they think that the market has a memory, that it will remember it had some bad years and make up for it in the future, in order to return to the “normal” long-run average.

Well, sorry kids, it just don’t work that way. The 10% number may still be sound as an annual estimate. Assuming it is, then 17 years at 10% compounds to 405.44%. Starting with $81.88, the expected value on January 1, 2026 is now $413.86. Sorry ’bout that.

No Comments

No comments yet.

RSS feed for comments on this post. TrackBack URI

Leave a comment

WordPress Themes